镜像站点 · 本页由第三方 GitHub 只读镜像提供,非 GitHub 官方站点,不接受任何登录或凭据输入。前往 github.com
Skip to content

Does even bubbling exist, or how to enable cancelsTouchesInView ? #30

Description

@jlongster

I don't see a stopPropagation method on the Event instance we get in a touch handler. Is there a way to do the equivalent of setting cancelsTouchesInView = YES on a view so that touch events do not go to parent views?

We probably don't need something like stopPropagation because you can't really do that on iOS, but some static property to set on a View.

Activity

  1. vjeux commented on Feb 2, 2015

    @vjeux
    Contributor

    Can you describe your use case? Hard to give a good answer without knowing what you are trying to do

  2. jlongster commented on Feb 2, 2015

    @jlongster
    Author

    I'm doing something a little unconventional, which you saw me tweet about: throwing a UI on top of a game. It's primitive right now but looks like this:

    screen shot 2015-02-02 at 11 22 16 am

    Those are up/down buttons at the bottom. Problem is, you can touch the screen to move around the camera, but if you press and drag your finger on top of those buttons, it still moves the camera around.

    My views are set up like this: one controller that has a normal UIView, the GL view as the first subview and the RCTRootView as the second subview.

  3. ericvicenti commented on Feb 2, 2015

    @ericvicenti
    Contributor

    We have a responder system to handle this, but we didn't have much time to explain how it works at the conference. There are a number of handles that we use to negotiate touch handling. These are working well on View, but not everything has full support for this yet. (I'm not sure what 3D view you're using here)

    A view can become the touch responder by implementing the correct negotiation methods. There are two methods to ask the view if it wants to become responder:

    View.props.onStartShouldSetResponder: (evt) => true, - Does this view want to become responder on the first touch-down?
    View.props.onMoveShouldSetResponder: (evt) => true, - Called for every touch move on the View when not responder: does this view want to "claim" touch responsiveness?

    If the View returns true and attempts to become the responder, one of the following will happen:

    View.props.onResponderGrant: (evt) => {} - The View is now responding for touch events. This would be a great time to highlight and show the user what is happening
    View.props.onResponderReject: (evt) => {} - Something else is the responder right now and will not release it

    If the view is responding, the following handlers are used:

    View.props.onResponderMove: (moveEvt) => {} - The user is moving their finger
    View.props.onResponderRelease: (releaseEvt) => {} - Fired at the end of the touch, ie "touchUp"
    View.props.onResponderTerminationRequest: (evt) => true - Something else wants to become responder. Should this view release the responder? Returning true allows release
    View.props.onResponderTerminate: (evt) => {} - The responder has been taken from the View. Might be taken by other views after a call to onResponderTerminationRequest, or might be taken by the OS without asking (happens with control center/ notification center on iOS)

    There are a few more features and facets to this system, which can be found in the highly documented ResponderEventPlugin.js

  4. jlongster commented on Feb 2, 2015

    @jlongster
    Author

    Thanks for the explanation, @ericvicenti! After I posted this I did see the diagram in ResponderEventPlugin.js.

    I believe, however, that is doesn't fix my use case. All of this is only happening within the views controlled by ReactKit, correct? I'm rendering everything from ReactKit into it's own subview, and rendering everything from OpenGLES into a different subview. So basically I can't control how I interact with anything outside of the ReactKit subview, right? By default, all touches are being handled by both of these views.

    The right way to do it, it seems, is to implement my own custom GL view on the native side and expose that as a custom component. Then I can render the GL view within my component hierarchy, and I can control the interactions.

    Where should I start to look at implementing my own native UIView subclass?

  5. vjeux commented on Feb 2, 2015

    @vjeux
    Contributor
  6. ericvicenti commented on Feb 2, 2015

    @ericvicenti
    Contributor

    Correct- responder negotiation happens in JS. The 3D view should be expressed in your React component hierarchy to use the responder system. (There currently isn't good responder negotiation between the top-level ReactKit component and external iOS components.)

    I'll defer to @jordwalke, @sahrens, @frantic, and @nicklockwood to help explain the state-of-the-art when it comes to bridging new components.

  7. jlongster commented on Feb 2, 2015

    @jlongster
    Author

    I think I see how. I need to create a component with a viewConfig property, which has a uiViewClassName property, which is the Obj-C class that it will instantiate when creating the backing iOS instance. And I need to implement my custom UIView class on the iOS side with this same class name. Theoretically it will instantiate my custom view class.

    I'll take it from here and file an issues I run into.

  8. jordwalke commented on Feb 2, 2015

    @jordwalke

    @jlongster Yes, that should work, but I don't know why your other view is receiving touches that started on the buttons. Are your touch handlers for the camera on the outermost view that contains everything? If so, you could put your touch handlers (or gesture recognizers) on just the part of the screen that has the teapot.

  9. jlongster commented on Feb 2, 2015

    @jlongster
    Author

    @jordwalke The 3d scene is actually rendered the full width/height of the phone, and the buttons are sitting on top of it. I do it this way because the UI will frequently change on top of the 3d scene; it's basically just a 2d overlay.

    I could shrink the 3d scene, but there will be many other cases where I want an input on top of the scene.

  10. jordwalke commented on Feb 3, 2015

    @jordwalke

    I see. Then your approach of creating a custom 3d-scene component sounds good. (BTW, we're going to streamline making new bridged components, and we're also going to work on integrating the JS cross-platform responder system with each platform's gestures).

    Also, where are your touch handlers? I think there's a way to make this work simply by making the buttons not under the place where your touch handlers are setup in iOS (regardless of geometry/how much space they take up).

  11. jlongster commented on Feb 3, 2015

    @jlongster
    Author

    Cool, I assumed the bridge was early work and that it'll be easier to make custom components in the future.

    The touch handlers are on the view controller controlling the GL view. But I just realized that is the root view of the window. I add the RCTRootView as a subview on top of my GL view, but you're right, my touch handlers are still on the controller of the whole thing. I probably could try just having multiple controllers and make sure the UI is sitting on top of everything.

  12. jordwalke commented on Feb 3, 2015

    @jordwalke

    @jlongster that sounds like the right approach (make sure the UI is really sitting on top of everything).

  13. added a commit that references this issue on Apr 18, 2018
    75db9fd
  14. locked as resolved and limited conversation to collaborators on May 29, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions